DeepSeek V3.1-Terminus

DeepSeek·Operational·Released 2025-09-22text
Context128k
Max output-
Input / 1M-
Output / 1M-
Cached in-
Benchmarks12

Benchmarks

Cited public sources · rank is exact-benchmark, apples-to-apples

LMArena Elo#88/1411417.2 ↗
Aider-Polyglot#1/876.1% ↗
BrowseComp#21/2338.5% ↗
BrowseComp-zh#3/345% ↗
GPQA-Diamond#4/880.7% ↗
LiveCodeBench#2/1174.9% ↗
MMLU-Pro#10/4985% ↗
SimpleQA (agentic tool use)#2/296.8% ↗
SWE Verified#1/268.4% ↗
Terminal-bench#3/336.7% ↗

DeepSeek V3.1-Terminus FAQ

What is DeepSeek V3.1-Terminus?
DeepSeek V3.1-Terminus is a large language model from DeepSeek, released on September 22, 2025. It accepts text input.
When was DeepSeek V3.1-Terminus released?
DeepSeek released DeepSeek V3.1-Terminus on September 22, 2025.
What is DeepSeek V3.1-Terminus's context window?
DeepSeek V3.1-Terminus has a 128K-token context window.
How does DeepSeek V3.1-Terminus score on benchmarks?
DeepSeek V3.1-Terminus has 12 cited benchmark scores, including LMArena Elo 1417.2 Elo, Aider-Polyglot 76.1% and BrowseComp 38.5%. Each score links to its source.

Access DeepSeek V3.1-Terminus and every other model through one endpoint with automatic failover: Respan gateway.